MCP-based E-commerce Analytics Chatbot

This project is an end-to-end analytics solution for e-commerce platforms, featuring:

1. Backend MCP Server (server.py)

   • Exposes tools to fetch orders, invoices, attendance, sales metrics, and user sales details via MCP protocol.
   • Connects to multiple customer databases using credentials from environment variables.
   • Handles robust error reporting and reconnection logic.

2. Chatbot Client (chatbot.py)

   • Built with Streamlit for an interactive web-based chat interface.
   • Supports both OpenAI and Google Gemini LLM providers (select via the LLM_PROVIDER env var).
   • Parses user queries, extracts intent/parameters, and calls the backend tools.
   • Dynamic formatting for all MCP tools and improved error handling.

---

Setup & Running the Project

1. Prerequisites

• Python 3.9 or higher
• Access to your e-commerce Postgres databases
• API key for your preferred LLM provider (OpenAI or Gemini)

2. Installation

1. Clone the repository:

   git clone <repo-url>
   cd rapidor-mcp
   

2. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure environment variables:
   • Copy .env.example to .env and set:
     • Database: MASTER_DB_NAME, MASTER_DB_USER, MASTER_DB_PASSWORD, MASTER_DB_HOST, MASTER_DB_PORT, INITIALIZED_DBS
     • LLM: OPENAI_API_KEY or GEMINI_API_KEY
     • LLM selection: LLM_PROVIDER (set to openai or gemini)

A. Start the Streamlit Chatbot UI

streamlit run chatbot.py --server.port 8080

This launches a browser-based chat interface (typically at http://localhost:8501).

Usage

• Ask analytics questions in natural language (e.g., "Show total sales for john in acmeco.com from 2025-01-01 to 2025-01-31").
• The chatbot will extract intent, call the backend, and present the results.
• Supports queries for orders, invoices, attendance, leave, total sales, and user sales details.

Example Queries

• How many orders did user 'antim' have in testdemo7.rapidor.co between 2022-03-01 and 2022-03-10?
• Show me total sales for user 'john' in domain 'acmeco.com' from 2025-01-01 to 2025-01-31.
• List attendance for user 'susan' in domain 'example.com' for May 2025.
• Show Users and their sales details in testrsfp.rapidor.co from 2025-01-01 to 2025-06-30.

---

Troubleshooting

• OpenAI/Gemini errors: Check your API key, quota, and LLM_PROVIDER setting.
• Database errors: Verify credentials, database status, and that INITIALIZED_DBS contains the correct DB names.
• Streamlit issues: Ensure the MCP server is running and that there are no port conflicts.
• No databases found: Check your .env configuration and that the backend server can access the master database.
• General errors: Review terminal output for detailed logs.

---

Contributing

Pull requests and issues are welcome! Please open an issue to discuss major changes first.

---

License

MIT License. See LICENSE file for details.

• Enter your analytics questions in natural language, specifying the user, domain, and date range as needed.
• The chatbot will parse your request, query the backend, and display the results instantly.

Example Queries

• How many orders did user 'antim' have in testdemo7.rapidor.co between 2022-03-01 and 2022-03-10?
• Show me total sales for user 'john' in domain 'acmeco.com' from 2025-01-01 to 2025-01-31.
• List attendance for user 'susan' in domain 'example.com' for May 2025.

---

Troubleshooting

• If you see OpenAI errors, check your API key and quota.
• For database errors, verify your credentials and database server status.
• If Streamlit does not load, ensure the MCP server is running and there are no port conflicts.

---

Contributing

Pull requests and issues are welcome! Please open an issue to discuss major changes first.

---

License

MIT License. See LICENSE file for details.